Skip to main content

Reply to this post | See parent post | Go Back
View Post [edit]

Poster: Kimo C Date: Nov 3, 2005 4:57am
Forum: sflan Subject: on a lighter note...

I think we can all agree that SPAM is bad, and that it is encouraging to see significant progress in combating it!

In fact i spoke recently with one of the foremost researchers in this field http://www.merl.com/people/yerazunis/ and he was open to requesting that when his home page is updated (not a high priority), that he not obscure his email in any way to prove how far things have come.


http://dspam.nuclearelephant.com/

The DSPAM (as in De-Spam) Project is a scalable, open-source statistical anti-spam filter. While most commercial solutions only claim a mere 95% accuracy (1 error in 20), a majority of DSPAM users frequently see around 99.95% (1 error in 2000) and can sometimes reach peaks as high as 99.991% (2 errors in 22,786, as with one particular user). DSPAM can be used as either a server-side filter (with many different configurations, including an SMTP gateway) or a developer's library (libdspam) for mail clients, other anti-spam tools, and similar projects requiring drop-in spam filtering. DSPAM has been implemented on many large and small scale systems with the largest being reported at about 350,000 mailboxes. It is presently being used or planned for use in multiple commercial spam filter solutions.

DSPAM is a dedicated statistical filter and in being such provides much higher levels of accuracy than commercial "hodge-podge" solutions, and with minimal resources. DSPAM's best recorded levels of accuracy have included 99.991% by one avid user (2 errors in 22,786) and 99.987% by the author (1 error in 7000), which is ten times more accurate than a human being! [1].